CÓDIGO CANTIDAD UD RESUMEN PRECIO SUBTOTAL IMPORTE C06.05.02 Ud Alimentación de agua potable, de 7,57 m de longitud, enterrada

It is generally assumed that κραταιός, the first member κραται-, and κραταιΐς (PN Kράταιϊς) had pre-forms starting with *kr̥t-. Although there is no reason to assume that -ρα- replaced an older full grade in these forms, their morphological analysis (especially the origin of -αι-) remains an object of debate to this date. In the present section, I will review previous attempts to reconstruct κραταιός and κραται-, and offer a new proposal for both forms; κραταιΐς and Kράταιϊς are discussed in the next section. Anticipating the conclusions to be reached in chapter 6, there is one strong indicator for the former presence of *r̥: of all lexemes containing the root κρατ-, the forms κραταιός, κραται-, and κραταιΐς are the only ones to undergo muta

cum liquida scansion.540 In chapter 6, I will argue that the outcome -ρα- in such forms is the

regular reflex of *r̥ in Epic Greek.

537 _{As I will argue in section 5.3, *kr̥teró- ‘impetuous, violent’ can be etymologically equated with Ved. śithirá-}

‘loose’ < PIE *ḱrth1-ró-.

538

As we have seen in section 5.2.1, such a semantic development also seems to be presupposed by the diverging semantics of the Cretan forms καρτερος ‘firm’ and καρτος ‘violence’.

539 _{Given the semantic difference between *krétu- ~ *kr̥téw- and *kr̥teró- reconstructed here for Proto-Greek, it}

remains to explain how κρατύς could be absorbed by κρατερός in Epic Greek. Could this be due to the same development reflected in the Ionic of Herodotus, where καρτερός retains the same wide gamma of meanings? This is not the only option: as so often, Herodotus could be Homerizing, and it remains unclear why Ionic would have given up the otherwise clear formal and semantic distinction between κρατύς and καρτερός. I am therefore inclined to consider an inner-Epic mechanism for the conflation of κρατύς and κρατερός ~ καρτερός. The form κρατύς is attested only in the formula |H κρατύς Ἀργεϊφόντης. When κρατύς ‘fierce’ had to qualify other names

like Diomedes in verse-final position, its use was inhibited by Hermann’s Bridge. In fact, κρατερός occurs after |H in almost 60% of its instances. It is conceivable that there was an old distribution between κρατύς and

κρατερός in pre-Homeric Epic: |H κρατύς |B V- [PN] ## versus |H κρατερός C- [PN] ##. After κρατύς had gone

out of use in spoken Ionic-Attic, its fate in Epic Greek was sealed by the much greater metrical utility of κρατερός, and it survived only in one single relic formula. It might then be speculated that formulae like |H

κρατερὸς ∆ιοµήδης and |H κρατερὴ ὑσµίνη date back to a stage when *kr̥teró- still had the meaning ‘impetuous,

fierce, violent’, and that *kr̥teró- underwent the development of Epic *r̥, to be discussed in the next chapter. If the semantic development of *kr̥teró- to ‘steadfast’ was indeed Proto-Greek, this would suggest a very high antiquity for the epic hexameter. Needless to say, this scenario remains uncertain, but it seems clear that κρατύς and κρατερός were somehow conflated in Epic Greek.

540 _{The muta cum liquida licence is never used in other forms with κρατ- (κρατύς, κράτος, κρατερός, κρατέω).}

As we have seen, κραταιός is semantically equivalent to κρατερός ~ καρτερός.541 Its reconstruction has been much debated, but a convincing solution has not yet been provided. It is often assumed that the masculine κραταιός is a back-formation from the feminine, attested in the Homeric formula Mοῖρα κραταιή.542 Authors like Risch, de Lamberterie, and Meissner suppose that κραταιή continues an archaic motional form *kr̥th2u̯ih2 of the u-stem adjective

κρατύς, where the second -a- would be the outcome of a vocalized *h2.543 This explanation is inspired by Πλάταια (Πλαταιαί), which is analyzed as the direct outcome of a motional feminine *pl̥th2u̯-ih2 ‘wide’ (cf. πλατύς).544 The reconstruction as a relic u-stem form is corroborated by the pattern of accentuation (singular Πλάταια, plural Πλαταιαί), which recurs in a few other archaic motional feminines of Greek u-stem adjectives (see section 4.1). Furthermore, the root-final *-h2- can be reconstructed on the basis of the Indo-Iranian cognates and of the Greek substantive πλαταµών ‘flat surface’.

When this explanation is extended to κραταιός, however, severe problems arise. First of all, there is no independent evidence that the root underlying κρατ- ended in *-h2. Meissner accepts the etymological connection with Vedic krátu- and Avestan xratu-, but these forms exclude a root-final laryngeal; he therefore assumes that a suffix *-h2u- (replacing older *-u-)

became productive in u-stem adjectives in Proto-Greek, which is clearly an ad hoc explanation.545 De Lamberterie (1990: 352-3) derives κρατύς from the root *kert- ‘to cut’, and is forced to assume a contamination between *kert- and *kerH- ‘to cut’, which would have given rise to *kerth2-. Both ideas are designed specifically in order to explain κραταιός, and neither of them is supported by further evidence.

A second problem concerns the inner-Greek developments that are assumed to lead to

κραταιός. The expected feminine of a u-stem adjective would be *κραταια (with short -ᾰ), but exception is Il. 20.121, where κράτος stands after the trochaeic caesura. For such incidental cases of muta cum liquida scansion, see section 6.4.

541 _{According to de Lamberterie (1990: 337), κραταιός “présente la même gamme d’emplois que la formation en}

-ερός, et cela est vrai aussi des composés en κραται˚, καρται˚.” To his representative list of examples I would add that the equivalence of κραταιός and κρατερός is most clearly proven by Od. 18.383, where κραταιός occurs in a speech of Odysseus (18.365-86). The “beggar” Odysseus addresses the suitor Eurymachos and warns him that Odysseus would beat him in any contest of endurance, be it in mowing the grass from morning till evening, in ploughing a field all day long, or in full war. This means that κραταιός, in the verse καὶ πού τις δοκέεις µέγας ἔµµεναι ἠδὲ κραταιός (18.383) “you think you are some big and tough guy”, refers to the stamina or endurance which the suitor Eurymachos, a daily consumer of banquets, is lacking. This shows that κραταιός does not only appear in the sense ‘impetuous, fierce’ (likely in its other Homeric instances, and probably the more original meaning), but that it also means ‘enduring’ on at least one occasion.

542 _{Risch (1974: 74), Nagy (1999: 85-89 and 349-54), de Lamberterie (1990: 337ff.), Meissner (2006: 62f.). In}

Risch’s words, “danach [i.e. after Πλάταια] wohl zu κρατύς : µοῖρα κραταιή (statt *-αιᾰ), Versende 9 mal Il. (nachträglich ist κραταιός gebildet worden).”

543 _{Most scholars reconstruct PIE *kr̥th}

2u̯-ih2, but departing from such a pre-form, I would expect the *u to

surface in Greek as a syllabic segment. For this reason, I prefer to reconstruct the Greek form as *kr̥th2-eu-ih2:

see the next note.

544 _{This toponym may originally have denoted a wide or flat area (e.g. *πλάταια χώρα). Of course, the normal}

feminine πλατεῖα to πλατύς is analogical. As explained in section 4.1.1, the reconstruction *plth2-u-ih2 required

by Ved. pr̥thivī- would not yield Greek Πλάταια, because one expects the *u to surface as a syllabic segment in the Greek outcome. This means that the Greek and Indo-Iranian forms cannot be derived directly from one and the same pre-form. In my view, the full grade suffix -ew- presupposed by the Greek feminine (*plth2-eu-ih2) was

taken from the weak stem of the masculine. Risch’s (1974: 74) suggestive reference to Pausanias, according to whom Πλάταια used to harbor a cult of the goddess Earth, does not necessarily imply a direct formal equation between Πλάταια and Vedic pr̥thivī- ‘Earth’.

545 _{Meissner uses this reconstruction to explain the compounding element κραται- from κραταιο-, which would}

have been altered under pressure of other compounds with a “linking element -αι- (of various origins) which is favoured over -ο- wherever possible and (…) thus found not only in κλυται-, πυλαι-, κελαι- for κελαινο-, etc. [in Κλυταιµνήστρη, Πυλαιµένης, κελαινεφής, LvB] but even in comparative forms like γεραίτερος alongside γεραιός.”

this form is unattested, and κραταιή is found in the archaic Homeric formula Mοῖρα

κραταιή.546 As explained in section 4.1, in a pre-form Ns. *κράταια I expect the accent to be on the root, which makes the final accent of κραταιή problematic. Furthermore, there is no apparent motive for replacing an earlier *Mοῖρακραταιά by Mοῖρακραταιή. The creation of a secondary masculine κραταιός beside κραταιή is not without problems either. De Lamberterie (1990: 339) proposed to call κραταιός a “masculinatif”, but most of the parallels cited for this process are from post-Classical Greek, and the similarities with κραταιός are only partial.547

In view of these objections, the proposal to derive κραταιός from the feminine of a u- stem adjective remains doubtful.548 Before further analyzing κραταιός, let us now consider the first member κραται-, which requires an explanation too. A number of details concerning its reconstruction remain unclear:

(1) What is the origin of -αι-? Is there a derivational relation between κραται- and

κραταιός?

(2) A first member κραται- was metrically awkward, because it entailed the use of

muta cum liquida scansion. Why was it created at all?

(3) What is the relation between names with Kραται- and those with Kαρτι-, Kρατι-? Let us first consider the existing opinions on the origin of -αι- in κραται-. Meissner argues that κραται- is a remodeling of κραταιο-, citing γεραιός ‘old’ beside the comparative

γεραίτερος as a parallel.549 This is unlikely: it remains unclear why the thematic vowel would be dropped in the first member of a compound. Moreover, the assumed influence of γεραιός on κραταιός lacks a semantic motivation. Finally, γεραίτερος is not a compound but a comparative, and the relation between γεραιός and γεραίτερος is unclear itself.

For κραται-, the reconstruction of a pre-form *ḱrth2i- does not really help: between two consonants, PIE *-h2i- is expected to yield -ι- rather than -αι- (cf. Nagy 1999: 86f. n. 5). Nagy assumes that κραται- arose within Greek as a cross of the adverb *kr̥ t-a with the

expected Caland allomorph *kr̥t-i-. He follows a suggestion by Nussbaum that adverbs in -α

could also appear in place of a first member in -ι-, as in the names Ἀλκάθοος (Il. 12.93) and

Ἀλκαµένης (Bechtel 1917: 35) beside ἀλκί-φρων. Indeed, a first member Kρατι- is also

546 _{De Lamberterie (1990: 340) proposed to recognize a trace of *κραταιά in κραταιΐς (Od. 11.597), which he}

derives from a syntagm *κραταιὰ ϝίς ‘strong force’ by haplology. But this scenario cannot be further substantiated; for a different analysis of κραταιΐς, see below.

547 _{A masculine form (ϝ)ιδυῖος, which refers to an overseer of youngsters, occurs in late Laconian and Messenian}

inscriptions from the 2nd and 1st c. BC. The same form recurs in glosses ascribing the form to Attic in the meaning ‘witness’ or ‘jury in a homicide case’. De Lamberterie proposes that (ϝ)ιδυῖος was based on the feminine pf. ptc. ἰδυῖα, which seems likely. It deserves attention, however, that (ϝ)ιδυῖος is a substantivization, whereas κραταιός is a full-fledged adjective. None of the other examples of “masculinatives” furnishes a clear parallel for κραταιός: beside πέπων (fem. πέπειρα), a new masculine form πέπειρος is found first in the Hippocratic corpus; beside Homeric θαµέες (fem. θαµειαί), the thematic comparative form θαµειότερος first occurs in Nicander. Finally, beside the inherited adjective στεῖρα ‘barren, sterile’, typically used with female referents, the secondary masculine στεῖρος is first attested in Euripides. Among the alleged examples, the only clear Homeric case is ἑταῖρος ‘companion’ beside ἕταρος. But here, it is hard to exclude influence of the feminine ἑταίρη on an already existing masculine form ἕταρος.

548 _{We have already seen that the u-stem form κρατύς may have been productively formed within Greek, induced}

by the semantic differentiation between the precursors of καρτερός and κράτος. Again, this casts doubts on the idea that κρατύς and its supposed feminine *κράταια existed when the root-final *-h2- was still intact. I will

argue below (section 5.3) that the inherited form of the adjective κρατερός ~ καρτερός corresponds directly to Vedic śithirá- ‘loose’. This identification does point to a root-final laryngeal, then, but the root must be set up as

*ḱreth1- in view of the -ε- in Ion.-Att. καρτερός. For the sake of the argument, one could be tempted to assume

that the root was *ḱreth2-, and that *καρταρός was reshaped to καρτερός within Greek, for instance after ἱερός,

which was originally close in meaning (‘active, agile’). However, this would not solve the other problems with previous explanations of κραταιός. It does not explain either why Cretan has καρτερος: corresponding to Hom. ἱερός, West Greek dialects have ἱαρός.

549

found, but only in personal names (Nagy 1999: 86). This scenario receives some support from the fact that an adverb *kr̥ta is attested in Classical Greek as κάρτα. One would have to assume that *kr̥ta-, of adverbial origin, could be used as the first member of a compound, and

then extended with the Caland suffix -i- or contaminated with *kr̥ti-. But unfortunately, there

are no clear parallels to support this scenario.

A model for the creation of *kr̥tai- is difficult to indicate. But if we assume that this

form is old, a motivation for its retention is available within Epic Greek. In compounds, the first member κραται- functions as an allomorph of κρατερός. This is clearly illustrated by personal names with Κραται- which correspond to Epic syntagms with κρατερός, cf. de Lamberterie (1990: 337):

Κραταιµένης (Th., inscr.)550 ~ Hom. κρατερὸνµένος ‘impetuous fury’

Κραταίβιος (inscr. Delos)551 ~ Hom. κρατερῆφιβιῆφι ‘with vigorous force’.

Another piece of evidence is καρταίποδ-, attested both in Pindar (qualifies a bull in Ol. 13.81) and in Cretan inscriptions in the meaning ‘cattle’. Furthermore, κραταίποδες ‘with strong feet’, epithet of ἡµίονοι ‘mules’, is attested in the Lives of Homer.552 This appellative compound must be compared primarily with the Homeric epithet κρατερῶνυξ, which (with one exception) qualifies horses and mules that are used as draught animals.553 In other words,

κραταίποδες is the equivalent of a phrase “whose πόδες are κρατεροί”, just like κρατερῶνυξ means “whose ὤνυχες are κρατεροί”.

Given that κραται- functions as an allomorph of κρατερο- in compounds, we have to examine their distribution. In my view, the underlying principle is metrical and must be sought in Epic Greek. The only two Homeric compounds with a first member κραται- are

κραταιγύαλοι ‘with solid breastplates’ (Il. 19.361, qualifies θώρηκες) and κραταίπεδον (Od. 23.46, qualifies οὖδας ‘ground’). In these cases, the second member starts with a single consonant and has a short first syllable. The same applies to the post-Homeric personal names

Κραταιµένης and Κραταίβιος, which are clearly of Epic origin, and to κραταίποδες (cf. Cretan and Pindaric καρταίποδ-). In this context, both κρατερο- and καρτερο- were excluded for metrical reasons in Epic Greek, because they would have yielded a sequence of three or more short syllables. In front of a second member starting with two consonants or with a vowel- initial heavy syllable, we find κρατερο-: cf. κρατερόφρων, κρατερῶνυξ. When the first syllable of the second member was heavy and started with a single consonant, καρτερο- was used: cf. καρτερόθῡµος). Thus, the distribution is as follows:

First member: Second member starts with:

καρτερό-(θῡµος) CV̄ -, CVCC-554

κρατερό-(φρων) CC-

κρατερ-(ῶνυξ) V̄ -, VCC

κραταί-(πεδον) CV̆ (C)V̆ -

550 _{The oldest attestation is the name of an Achaean victor in Olympia (SEG 22.345, appr. 600 BC). Further}

attested (mostly late) in Κραταιµένου SEG 19.108 I.117 (Attica, cf. SEG 23.124.2), Κραταιµένης IG V(1) 127.4 and 211 II.34 (Laconian), also IG V(2) 419.8 (Arcadian, 2nd c.), Εὔδ]ηµος Κραταιµένου Ἐρετριεύς IG XII(9) 91.4 (Euboea). Cf. also Καρταιµ̣έ̣νη[ς], IG XII (Supp.) 312 III.31 (Tenos, Ionic Cyclades, 2nd _c.).

551 _{Κραταίβιος IG XI(2) 287 A.146 passim (Delos). The form Καρταίβιος (with -αρ-) is also attested as the name}

of a Cretan in Miletus (Bechtel 1917: 256).

552 _{Vit. 19, 4 (ed. Wilamowitz).}

553 _{The exception is λύκοι κρατερώνυχες ἠδὲ λέοντες (Od. 10.218), ‘with violent claws’ (vel sim.).}

554 _{In κραταιρίνοιο ‘hard-shelled’ (oracle in Hdt. 1.47, hexameter), the allomorph κραται- is used in front of a}

The compounds with κραται- were preserved because they could not be replaced by compounds with *kr̥tero- within Epic Greek.555 This makes it likely that κραται- represents a relic “Caland allomorph” of *kr̥tero- as a first member.

The question then arises how the compounds with κραται- relate to personal names with Kαρτι- and Kρατι-, which contain the expected outcome of a pre-form *kr̥ti-. As

evidence for such names, Bechtel (1917: 256) mentions Kρατ-ερµος, Kρατ-ιππιδας, Kρατι-

δηµος and Kαρτι-δαµας, Kαρτι-νικος, Kαρτι-σθενης. Meissner (1998: 244-45, cf. also Frisk s.v. κράτος) remarks that the attestations are not very early.556 For this reason, he claims that these names could be innovations of the classical period, when first members with -ι- enjoyed productivity, and remarks (1998: 245): “das Fehlen von κρατι- bei Homer [ist] nicht auf metrische Gründe zurückführbar. (…) Ein καρτι- bzw. κρατι- wäre metrisch vielseitig verwendbar. Sein vollständiges Fehlen ist also auffällig.” However, given that κραται- occurs only twice, the absence of καρτι- ~ κρατι- in Homer is not necessarily remarkable in the first place. Furthermore, the above distribution suggests that κραται- was preserved in Epic Greek merely because it could not replaced by κρατερο-. Since forms with *kr̥ti- may underlie the

forms with *kr̥ tero-, Meissner’s argument is invalid. Since κρατερο- ~ καρτερο- is unattested as the first member of PNs in Ionic-Attic (see section 5.2.2), it seems much more likely that the personal names with Kρατι- and Kαρτι- directly continue the inherited form *kr̥ti-, and

that *kr̥ ti- was replaced by *kr̥ tero- only in appellatives.557

We now arrive at the following scenario. In Epic Greek, the use of the inherited allomorph *kr̥ti- was problematic in front of a single consonant followed by a short

syllable.558 This may have initially been solved by a metrical lengthening to *kr̥tī-, e.g. *kr̥ti-

pedo- >> *kr̥tī-pedo-. Subsequently, the remaining cases of *kr̥ti- were replaced by *kr̥ tero-

whenever this was possible (or by κρατερο- ~ καρτερο-, as soon as this option became available). Finally, the isolated first member *kr̥tī- was replaced by *kr̥tai- (which eventually

developed to Hom. κραται-) under the influence of κραταιός. As we will see in chapter 6, this scenario accounts for the synchronically irregular muta cum liquida scansion of κραται-: the metrical lengthening occurred when *r̥ was still in place.

Let us again return to the reconstruction of κραταιός. In view of its oxytone accent, a derivation with the unaccented suffix -i(i̯)o- can be excluded. As possible parallel formations,

three other adjectives are of special interest: παλαιός ‘of the past’, γεραιός ‘old’, and δηναιός ‘long-lived’.559 Of these, δηναιός occupies a special position, because it probably derives

555

As a consequence, the compound κραταίλεως ‘consisting of hard rock’ (only attested in the tragedians, containing λᾶας ‘stone’ as its second member) must be considered a recent formation (note the Quantitative